Investigation on defect distribution and its statistical evaluation for case hardened material states

The fracture behaviour of steel components with very hard material states under cyclic loading is mostly defined by non-metallic inclusions present in the material. The inclusion distribution is a result of specific conditions in steel making, casting and rolling processes. Standard methods (e.g. DIN EN 10247:2007) are commonly used for estimation of the distribution of non-metallic inclusions. These methods are based on random metallographic examination of longitudinal microsections and evaluation with standard pictures. For numerical evaluation of the endurance limit of very hard material states the knowledge of defect distribution of the material is necessary. However, the standard methods are not suitable to find the actual inclusion distribution. In a large investigation program on the fatigue strength of very hard material states, thin wall tubes made of a case hardening steel were tested. All fracture surfaces were investigated, size and coordinates of the fracture initiating defects were documented and analysed statistically. The defects appear as inclusion stringers stretched in rolling direction mainly containing Al-oxide. Several distribution functions for statistical analysis were investigated. Furthermore, transverse microsections with a total area of about 2000 mm2 were investigated and the data of the maximum width of inclusion stringers were also analysed statistically. Comparison of the extreme value distribution of inclusions made by metallography and fractography, respectively, show that the metallographic method considerable underestimates the actual size of inclusions. Therefore micrographic methods are not suitable to generate input data for inclusion distribution based on fatigue design.